Maturing peptides open for communication

Two decades after the discovery of systemin, the first plant peptide shown to be involved in cell to cell communication (Pearce et al., 1991), plant peptide research is coming of age, and is ready for cross-talk with older research fields in plant science. This issue of JXB illustrates that plant peptide signalling is of crucial importance for all aspects of plant growth and development. Plants respond to peptides produced by pathogens, they use peptide signalling to reinforce defence responses, and peptides exert signalling functions in both above- and below-ground organs, in vegetative and reproductive tissues. The PEPTIDE issue presents the initial characterization of recently identified peptide families, and highlights current knowledge on receptor interactions, downstream signalling modules, feedback loops, and target genes for a selection of exogenous and endogenous peptides involved in defence and development. With the overall aim of understanding the molecular mechanisms behind the biological processes in which peptide signalling is involved, plant peptide research follows three major routes: detection and functional characterization of new peptide families; genetic, molecular, and biochemical identification of receptors of peptide ligands and their signalling networks; and characterization of downstream factors and targets of peptide-controlled signalling pathways. I would like to acknowledge the first European Workshop on Peptide Signalling in Plants which was arranged in January this year in Oslo; the workshop reflected the current status of plant peptide research and many of the topics covered in this issue. Putative peptide ligands have been discovered by three approaches—by use of biochemistry, genetics, and bioinformatics (Butenko and Aalen, 2012). Bioactive compounds with promoting or inhibiting effects on differentiation or cell proliferation have been isolated from plant extracts, and genes encoding the corresponding peptides have thereafter been identified. Phytosulphokine (PSK), Plant Peptide Containing Sulphated Tyrosine (PSY1), the Rapid ALkalinization Factor (RALF), and Tracheary Element Differentiation Inhibitory Factor (TDIF) were discovered by this biochemical approach (Yang et al., 1999; Pearse et al., 2001; Ito et al., 2006; Amano et al., 2007), which has the advantage of direct identification of the mature peptide generated by post-translational processing and modifications of a precursor protein. Genes encoding peptides such as CLAVATA3 (CLV3), TAPETUM DETERMINANT1 (TPD1), EPIDERMAL PATTERNING FACTOR1 (EPF1), and INFLORESCENCE DEFICIENT IN ABSCISSION (IDA) have, on the other hand, been discovered due to their loss-of-function or gain-of-function phenotypes (Clark et al., 1995; Butenko et al., 2003; Yang et al., 2003; Hara et al., 2007), directly linking the gene in question to a given developmental process or morphological feature. Most recently a family of egg cell cysteine-rich peptides (CRPs) was identified as necessary for successful male–female gamete interaction in Arabidopsis thaliana (Sprunck et al., 2012). Peptides identified by biochemical and genetic approaches have served as founding members of in silico identified peptide families, such as CLV3/ENDOSPERM SURROUNDING REGION-RELATED (CLE) and IDA-LIKE (IDL) (Butenko et al., 2003; Jun et al., 2008). In the current issue of JXB, expression patterns and functions of EPFs and EPF-LIKE (EPFL), ROOT GROWTH FACTOR/GOLVEN/CLE-LIKE (RGF/GLV/CLEL), C-TERMINALLY ENCODED PEPTIDEs (CEPs), and AtPEP families are presented (Bartels et al., 2013; Delay et al., 2013; Fernandez et al., 2013; Richardson and Torii, 2013; Roberts et al., 2013; Uchida and Tasaka, 2013). The families of characterized peptides represent <10% of the estimated number of secreted peptide ligands. Genome-wide bioinformatic analyses have detected >800 Arabidopsis genes encoding small putatively secreted CRPs (Silverstein et al., 2007). The other major group of secreted peptides encompasses families of small proteins (<100 amino acids) with N-terminal secretion signal, variable region, and C-terminal family-specific conserved motifs representing the active mature peptide. The small genes encoding precursors of such peptides were largely overlooked during the initial annotation of the Arabidopsis genome, but later bioinformatics searches have estimated the number of secreted putative peptide ligands at ~1000 (Lease and Walker, 2006). Protocols have been developed for bioinformatical detection of unannotated peptide genes based on general features of secreted peptides, and identification of members of peptide gene families using a known peptide as a starting point (Lease and Walker, 2010). New methodologies and approaches can be employed to confirm that postulated or unannotated genes are transcribed and encode active peptide signalling molecules. One example is the identification of short transcriptionally active regions (TARs) encoding novel oxidative stress-induced peptides, by using mRNA from paraquat-treated Arabidopsis plants in genomic tilling array hybridization (De Coninck et al., 2013). Next-generation sequencing of transcriptomes would similarly allow identification of transcriptional activity from postulated peptide genes and novel peptide-encoding short open reading frames found in intergenic regions. Biochemical and/or genetic evidence has been provided for less than a dozen plant peptide ligand–receptor pairs (Butenko et al., 2009); most recently, CLE45, involved in differentiation of root phloem cells, and genetically dependent on BARELY ANY MERISTEM3 (BAM3) (Depuydt et al., 2013). In the current issue of JXB, attention is given to CLE peptides and their receptors involved in maintenance of the root apical meristem in Arabidopsis, and a corresponding CLE peptide in rice (Chu et al., 2013; Stahl and Simon, 2013); IDA and the receptors HAESA (HAE) and HAE-LIKE2 (HSL2) controlling cell separation during both floral organ abscission and lateral root emergence (Aalen et al., 2013; Liu et al., 2013); EPFs and EPF-LIKE peptides that signal through the ERECTA (ER) and ER-LIKE (ERL) receptors to regulate stomatal development (Richardson and Torii, 2013); as well as the endogenous danger peptides AtPEPs that signal through the PEP RECEPTOR1 (PEPR1) and PEPR2 (Bartels et al., 2013); and EF-Tu RECEPTOR (EFR) and FLAGELLIN SENSING2 (FLS2) responding to the bacterial peptides EF-Tu and flagellin22, respectively (Albert, 2013). So far, receptors have not been identified for any of the RGF/GLV/CLEL or CEP peptides. How can we match more ligands with receptors? Ideally one would like to see mutations in peptide and receptor genes that generate the same phenotype (Butenko et al., 2012). This is, however, often obscured by functional redundancies of paralogous genes. The expression patterns of the peptide and receptor genes should be compatible with the assumed short-range movement of peptides in the apoplastic space. It is very useful to have an easily observable gain-of-function phenotype through overexpression and/or application of a synthetic peptide since this provides a means for genetic identification of factors, including receptors, necessary for peptide signalling. Phenotypes resulting from overexpression or exogenous peptides must be interpreted with caution, since they may result from interaction with non-native receptors, and genetic evidence for interactions should be supported by biochemical confirmation of direct binding of the peptide to its receptor. Nicotiana benthamiana has been used successfully in peptide–receptor activation and binding studies of MAMPs (microbe-associated molecular patterns) triggering signalling cascades of defence reactions when recognized by defence-related plant receptors. In this field, ‘address–message’ is the working hypothesis, meaning that the ligand–receptor interaction occurs in two steps—recognition and activation (Albert, 2013). Thus removal of amino acids involved in activation—for instance the six C-terminal amino acids of the 18 amino acid long elf18 peptide from Escherichia coli, which is recognized by the Arabidopsis EFR receptor—converts an active peptide into an antagonist that only binds, but does not elicit the receptor. Removal of the N- or C-terminal residue of the mature CLV3 (mCLV3) rendered the peptide inactive (Kondo et al., 2006), indicating that correct processing of the peptide at both ends is important. Dominant-negative versions of CLV3 have recently been generated by substituting the almost invariant glycine with a threonine in the very middle of the CLE motif (Song et al., 2013), while comparison of the CLE peptides, including CLV3, that bind the closely related BAM1 and CLV1 receptors (Shinohara et al., 2012), with those that bind the PHLOEM INTERCALATED WITH XYLEM/TDIF RECEPTOR (PXY/TDR1) receptor and are involved in xylem differentiation (Etchells and Turner, 2010), suggests that binding specificity is found at the N- and C-termini (Kondo et al., 2008). In CRPs, 4–8 conserved cysteines contribute to the three-dimensional structure by establishing disulphide bridges (Ohki et al., 2011). The cysteine-rich EPF and the EPFL peptides are highly conserved, but with a flexible central loop region. Exchanging this variable part of STOMAGEN/EPFL9, which functions as a positive regulator of stomatal density, with that of the negative regulator EPF2 swapped the effect of these peptides (Ohki et al., 2011), showing that the functional specificity resides in this region, which is well conserved across species. The >400 receptor-like kinases (RLKs) of Arabidopsis have been classified according to their kinase domains (Shiu and Bleecker, 2001). The receptors shown to date to interact physically with exogenous or endogenous peptide ligands belong to four out of 13 classes with extracellular domains consisting of leucine-rich repeats (LRRs): X (EXESS MICROSPOROCYTES1, EMS1), XI (CLV1, BAM1, TDF/PXY, PSKR1, PSKR2, AtPEP1, AtPEP2), XII (ERF, FLS2), and XIII (ER, ERL). These RLKs all have >20 LRRs, suggesting that a large structure is required for ligand binding and activation. The regions involved in binding are atypical compared with the neighbouring consensus LRRs (Shinohara et al., 2007, 2012; Jia et al., 2008; Lee et al., 2012). Amino acid sequence alignments to orthologues of a given LRR-RLK can provide valuable information on receptor-specific evolutionarily conserved residues that potentially are involved in ligand binding (Helft et al., 2011). For peptides generated from C-terminally conserved motifs, post-translational modifications, such as hydroxylation of prolines, glycosylation of hydroxyprolines, and sulphation of tyrosines, have been shown to be important for activity and binding (Matsubayashi, 2011). A tyrosylprotein sulphtransylase enzyme has been identified as crucial for sulphated peptide functions (Komori et al., 2009). Chemical synthesis of peptides with hydroxyprolines, and recently also glycosylation, together with methods for isolation of peptides in planta will be very useful in future peptide research (Matsubayashi, 2011; Shinohara and Matsubayashi, 2013). So far, processing enzymes releasing C-terminal peptides from precursor proteins have not been identified; although in vitro a protein extract from cauliflower digested CLV3 N-terminal to the position 1 arginine of the conserved CLE motif, while mass spectrometry analyses suggested the action of a progressive carboxypeptidase at the C-terminal side (Ni et al., 2011). Interestingly, the cauliflower extract could also digest IDA, and excess CLV3 partly inhibited this processing (Stenvik et al., 2008). Where and when does processing take place? The discovery of cross-talk between the IDA–HAE/HSL2 signalling pathway and the ADP-ribosylation factor GTPase-activating protein (ARF-GAP) NEVERSHED, that potentially regulate the trafficking of signalling molecules located at the plasma membrane or within the apoplastic space (Liu et al., 2013), emphasizes the importance of understanding the cell biology of peptide signalling. CLV3 and CLE40, involved in meristem maintenance in the shoot and the root, respectively, exemplify that slightly different peptides may signal through the same receptor, in this case CLV1 (Stahl and Simon, 2013). In the root meristem, CLV1 interacts with the small RLK ARABIDOPSIS CRINKLY4 (ACR4), and this complex is suggested to gate plasmodesmata between the quiescent centre (QC) and the columella stem cells. A novel concept is presented where peptide signalling controls the mobility of factors through the plasmodesmata rather than the activation of target genes. The concentration of CLE40 in the apoplastic space will determine to what degree the receptor complex is activated, which will result in differentiation of columella cells (Stahl and Simon, 2013). Current data indicate that there is not a one-to-one relationship between peptide ligands and receptors, and that related peptides can signal through related receptors. Recent examples are the eight peptides of the AtPEP family, representing endogenous danger and damage-associated molecular patterns (DAMPs) (Bartels et al., 2013). They induce pattern-triggered immunity (PTI) in a manner dependent on the highly similar receptors PEPR1 and 2. Overexpression of IDA results in HAE/HSL2-dependent early and ectopic abscission, and overexpression of each of the five IDL peptides gives a similar phenotypic outcome, suggesting that they can signal through these or closely related receptors when the spatial expression patterns are compatible (Stenvik et al., 2008). Exogenous application or overexpression of a number of CLE peptides provokes consumption of the shoot or root meristem in a CLV1- and/or CLV2-dependent manner (Meng et al., 2010), and several CLE peptides have been reported to bind these receptors as well as BAM1 and BAM2 (Guo et al., 2010). EPF1 and EPF2 ensure that individual stomata are separated by at least one cell. Both were recently reported to interact physically with ER, ERL1, and ERL2 when co-expressed in N. benthamiana leaves; however, in Arabidopsis leaves, EPF1 has a preference for ERL1, and EPF2 for ER (Lee et al., 2012). It is suggested that co-receptor complexes may modulate the receptors’ affinities for the peptides (reviewed in Richardson and Torii, 2013). To understand the outcome of peptide signalling in networks of ligands, receptors, and co-receptors it is therefore crucial to investigate their temporal and spatial expression patterns. Similar peptide–receptor modules may operate in different tissues, as illustrated by EPFL peptides: STOMAGEN/EPFL9, together with ER, is acting as a positive regulator of stomatal density in leaves; EPFL4/6–ER in stem elongation; while EPFL4/6–ERL1 is involved in procambial maintenance, and the ER family receptors also regulate floral patterning (Bemis et al., 2013; Richardson and Torii, 2013; Uchida and Tasaka, 2013). Moreover, the IDA peptide and its receptors HAE and HSL2 are involved in cell separation during both floral organ abscission and lateral root emergence (Aalen et al., 2013). Interestingly, these LRR-RLKs seem fully co-expressed and redundant in function in the abscission zone, but only partly in the layers overlaying the developing lateral root primordia. RLK EVERSHED (EVR) and SOMATIC EMBRYOGENESIS RECEPTOR KINASE1 (SERK1) are also involved in regulation of abscission, possibly by direct interaction with HAE and/or HSL2 (Liljegren, 2012; Liu et al., 2013). EVR and SERK1 have small extracellular domains with only five LRRs, which is also the case for SERK3, also known as BRASSINOSTEROID INDEPENDENT1 (BRI1)-ASSOCIATED RECEPTOR KINASE1 (BAK1), involved in signalling pathways controlling development as well as defence, through interaction with BRI1 and FLS2 receptors (reviewed in Li et al., 2012). BAK1 and BAK1-INTERACTING RECEPTOR-LIKE KINASE (BIR1) control a signalling pathway leading to cell death and the expression of defence-related genes. Interestingly, EVR has also been identified as SUPPRESSOR OF BIR1 (SOBIR) (Gao et al., 2009), suggesting that development and defence pathways share factors. Intriguingly, several components of mitogen-activated protein kinase (MAPK) cascades seem to be involved in several different pathways, for instance the stress-induced MAPKs MPK3 and MPK6 triggered by both AtPEPs and IDA (Cho et al., 2008; Bartels et al., 2013), raising the question of how peptide-specific signalling outcomes can be achieved. In this issue of JXB, such outcomes, for example expression and activation of transcription factors, modulation of auxin transport, and secretion of cell wall remodelling enzymes, are reviewed by Czyzewicz et al. (2013). The PEPTIDE issue of JXB illustrates that the peptide signalling field is providing insights into communication between plants’ cells for all aspects of plant biology, in both development and defence. Plant peptide signalling is of crucial importance for regulating cell division, cell expansion, cell differentiation, cell communication, cell separation, and cell death. However, the vast majority of peptide ligands, their receptors, and downstream components remain to be discovered and characterized. The approaches and concepts presented at the Oslo Workshop and the current issue of JXB may serve as inspiring examples for elucidation of the biological function and molecular properties of uncharacterized members of known peptide families as well as novel peptide families. Identification of peptides in species other than Arabidopsis, exemplified here by studies on CLE and CEP peptides of importance for root development in rice and Medicago truncatula, respectively (Chu et al., 2013; Imin et al., 2013), should be encouraged to facilitate research on peptide signalling from an evo–devo perspective. The peptide research field is maturing, and open for communication with other research fields in plant science.